Pair-Production of W Bosons in $e^+ e^-$ Interactions at $\sqrt{s}$ = 161 GeV

HAL Id: in2p3-00000233

http://hal.in2p3.fr/in2p3-00000233

Submitted on 16 Nov 1998

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Pair-Production of W Bosons in e ⁺ e ⁻ Interactions at √ s

= 161 GeV

M. Acciarri, O. Adriani, M. Aguilar-Benitez, S. Ahlen, B. Alpat, J. Alcaraz, G. Alemanni, J. Allaby, A. Aloisio, G. Alverson, et al.

To cite this version:

M. Acciarri, O. Adriani, M. Aguilar-Benitez, S. Ahlen, B. Alpat, et al.. Pair-Production of W Bosons in e

e

Interactions at √

s = 161 GeV. Physics Letters B, Elsevier, 1997, 398, pp.223-238. �in2p3-

00000233�

CERN-PPE/97-14 February 3, 1997

Pair-Production of W Bosons in e

e Interactions at

= 161 GeV

The L3 Collaboration

Abstract

We report on the measurement of W-boson pair-production with the L3 detector at LEP at a centre-of-mass energy of 161.34 GeV. In a data sample corresponding to a total luminosity of 11 pb

, we select four-fermion events with high invari- ant masses of pairs of hadronic jets or leptons. Combining all nal states, the measured total cross section for W-pair production is:

= 2

89

^: _:

(

)

0

14 (

) pb. Within the Standard Model, this corresponds to a mass of the W boson of: M

= 80

80

^: _:

(

)

0

03 (LEP) GeV. Limits on anomalous triple- vector-boson couplings are derived.

Submitted to Phys. Lett. B

1 Introduction

In the rst half of the 1996 data taking period, the e

e collider LEP at CERN was operated at a centre-of-mass energy,

, of 161.34 GeV. This centre-of-mass energy coincides with the kinematic threshold of the process e

e

W

W , thus allowing for the rst time the pair- production of W

bosons in e

e interactions. During this run the L3 detector collected a total integrated luminosity of 11 pb

.

To lowest order, three Feynman diagrams contribute to W-pair production, the

-channel

and Z-boson exchange and the

-channel

e exchange [1], referred to as CC03 [2{4]. The W boson decays into a quark-antiquark pair, for example W

ud or cs, or a lepton-antilepton pair, W

` , in the following denoted as

and

for both W

and W decays. In this article, we report on measurements of all four-fermion nal states mediated by W-pair production:

1. e

e

e

(

) 2. e

e

(

) 3. e

e

(

) 4. e

e

(

) 5. e

e

(

),

where (

) indicates the possible presence of radiative photons.

Additional contributions to the production of these four-fermion nal states arise from other neutral-current (NC) or charged-current (CC) Feynman diagrams. For high invariant masses of pairs of fermions and for the visible fermions all within the acceptance of the detector, the additional contributions are small. At the current level of statistical accuracy they need to be taken into account only for e

e

e

(

) (CC20) and e

e

(

) (CC56+NC56) [2{4].

The cross-section measurements for the ve signal processes are combined to derive the total cross section for W-pair production.

At threshold, these cross sections depend strongly on the centre-of-mass energy and the mass of the W boson, M

:

=

(M

). From the cross sections as predicted by the Standard Model for this centre-of-mass energy a value for M

is derived. The

-channel contributions to the cross sections contain the triple-vector-boson vertices

WW and ZWW.

Using the independent measurement of the W-boson mass at pp colliders [5], the total cross- section measurements allow us to set limits on anomalous triple-vector-boson couplings.

2 The L3 Detector

The L3 detector [6] consists of a silicon micro-strip detector [7], a central tracking chamber, a high-resolution electromagnetic calorimeter composed of BGO crystals, a lead-scintillator ring calorimeter at low polar angles [8], a scintillation counter system, a uranium hadron calorimeter with proportional wire chamber readout, and an accurate muon chamber system. A forward- backward muon detection system extends the polar angle coverage of the muon chambers down to 24 degrees in the forward-backward region [9]. These detectors are installed in a 12 m diameter magnet which provides a solenoidal eld of 0.5 T and a toroidal eld of 1.2 T. The luminosity is measured using BGO calorimeters [10] situated on each side of the detector.

The response of the L3 detector is modelled with the GEANT [11] detector simulation program which includes the eects of energy loss, multiple scattering and showering in the detector materials and in the beam pipe.

2

3 Measurement of Four-Fermion Production

The analyses described below reconstruct the four-fermion nal states. Charged leptons are explicitly identied using their characteristic signature. Hadronic jets are reconstructed by combining calorimetric energy depositions using the Durham jet algorithm [12]. Calorimetric clusters are treated as massless and are combined adding their four-momenta. The momentum of the neutrino in

events is identied with the missing momentum vector.

Selection eciencies and background contaminations of all processes are determined by Monte Carlo simulations. The following Monte Carlo event generators are used to simulate the various signal and background reactions: KORALW [13] (e

e

WW

(

));

EXCALIBUR [14] (e

e

(

)); PYTHIA [15] (e

e

(

)

ZZ(

), hadronic two- photon collisions); KORALZ [16] (e

e

(

)

(

)); BHAGENE3 [17] (e

e

e

e (

)).

Systematic errors on the cross-section measurements are conservative estimates and in all cases small compared to the statistical error. The measurement of the total luminosity,

, follows the procedure described in [18,19]. The total error on the luminosity measurement is estimated to be 0.6% [19].

The results on cross sections and couplings are determined in a combined t as discussed in Section 4.

3.1 e +

e !qqe()

Event Selection

The event selection for the process e

e

e

(

) requires an identied electron, missing momentum due to the neutrino, and high multiplicity arising from the

system. A

e

event selected in the data is shown in Figure 1.

The electron is identied in the electromagnetic calorimeter as the highest energy depo- sition with electromagnetic shower shape. This calorimetric cluster must have a polar angle of

cos

0

90 and an energy

larger than 25 GeV. In order to reject radiative photons, a track in the central tracking chamber must match the electron cluster within 10 mrad in azimuth. Electrons arising from decays of hadrons are rejected by requiring the electron to be isolated from the hadronic system. Isolation is imposed by asking that the electron energy is at least 70% of the total calorimetric energy deposited in a cone of half opening angle 15 degrees around the electron direction.

The neutrino energy

, inferred from the missing momentum of the event, must be larger than 25 GeV. In order to reject radiative

(

) events where the photon escapes along the beam pipe, the polar angle of the missing momentum vector must point well inside the detector,

j

cos

0

90.

The hadronic system is characterised by a large particle multiplicity. Requiring at least 15 calorimetric clusters rejects all purely leptonic nal states. After having removed the calori- metric energy depositions associated with the identied electron, the remaining calorimetric clusters are grouped into two jets. The masses of the two W bosons are calculated as the in- variant masses of the electron-neutrino system,

, and the jet-jet system,

qq . Both invariant masses are required to be larger than 50 GeV.

The distributions of the electron energy and of the invariant mass of the electron-neutrino system are shown in Figure 2, comparing data to Monte Carlo.

3

The above cuts select four events in the data. The selection eciencies and the background contributions are listed in Table 1. The signal eciency is determined within the following cuts:

E

e

;E

25 GeV;

cos

cos

0

90;

qq

50 GeV. The accepted background cross section is dominated by the processes e

e

(

) and e

e

e

e when one of the leptons escapes detection.

Systematic errors in the electron identication are derived from a comparison of data versus Monte Carlo using e

e

e

e (

) events as a control sample. Systematic errors on eciencies and accepted background cross sections are derived by comparing dierent Monte Carlo event generators and Monte Carlo samples simulated with dierent W masses and detector energy scales. A total systematic error of 5% on the measured cross section of the reaction e

e

qq

e

(

) within the above cuts is assigned.

3.2 e +

e !qq()

Event Selection

The event selection for the process e

e

(

) requires an identied muon, missing momentum due to the neutrino, and high multiplicity arising from the

system.

The muon is identied in the muon spectrometer as the highest momentum track pointing back to the interaction vertex. It must have a momentum larger than 20 GeV. Muons arising from decays of hadrons are rejected by requiring the muon to have an angular separation of at least 15 degrees to both hadronic jets reconstructed as described below. In order to reject

events, any additional muon reconstructed in the muon chambers must have a momentum of less than 20 GeV.

The neutrino direction is inferred from the missing momentum direction of the event. In order to reject radiative

(

) events where the photon escapes along the beam pipe, the polar angle of the missing momentum vector must point inside the detector,

cos

0

95. Requiring at least 15 calorimetric clusters and at least ve tracks in the central tracking chamber rejects all purely leptonic nal states as well as cosmic-ray background.

The calorimetric clusters are grouped into two jets. The masses of the two W bosons are calculated as the invariant masses of the muon-neutrino system and the jet-jet system. The muon-neutrino invariant mass must be larger than 55 GeV, and the jet-jet invariant mass must be larger than 40 GeV and smaller than 120 GeV.

The distributions of the angle between the muon and the nearest jet and of the invariant mass of the muon-neutrino system are shown in Figure 3.

Cross Section

The above cuts select four events in the data. The selection eciencies and the background contributions are listed in Table 1. The accepted background cross section is dominated by the processes e

e

(

) and e

e

when one of the leptons escapes detection.

Systematic errors are evaluated as described above. A total systematic error of 5% on the measured cross section is assigned.

4

3.3 e e !qq()

Event Selection

The event selection for the process e

e

(

) is based on the identication of a tau jet in a hadronic event, combined with missing energy. The tau jet is identied as a low-energy electron or muon, or a low-multiplicity narrow jet, isolated from the rest of the event.

Events are selected on the basis of the nal-state particle multiplicity. Events must have more than 15 calorimetric clusters, rejecting low-multiplicity leptonic nal states. High-multi- plicity purely hadronic nal states are rejected by a cut in the two-dimensional plane spanned by the number of tracks reconstructed in the central tracking chamber and the number of calorimetric clusters.

Requirements on the missing energy and momentum are imposed. Signal events contain at least two neutrinos, resulting in missing momentum and reduced visible energy. In order to reject

(

) and

(

) events the dierence between the visible energy and the missing momentum must be less than 120 GeV. Requiring the longitudinal energy imbalance to be smaller than 30 GeV and the transverse energy imbalance to be larger than 5 GeV suppresses

q^q

(

) events with hard initial-state radiation.

The tau lepton is identied by its decay products. Electrons and muons are identied according to the lepton identication described above. If the lepton energy is larger than 5 GeV and the sum of the lepton energy and the missing momentum less than 65 GeV, the identied electron or muon is considered as the tau jet.

If no electrons or muons are found, geometrical jets are reconstructed based on clustering inside a cone of 15 degrees half-opening angle. At least three jets with an energy larger than 10 GeV are required. Out of the three most energetic jets the two most back-to-back jets are associated with the

system. The most energetic remaining jet is taken as the tau jet. The eciency of this tau jet identication for hadronic tau decays is 83%. In order to reduce the background coming from

e

(

) events with the electron not identied, events with the tau jet having more than 35 GeV of energy deposited in the electromagnetic calorimeter and less than 3 GeV in the hadron calorimeter are rejected. The background of

(

) events with the muon not identied in the muon chambers is reduced by rejecting events where the tau jet is compatible with a minimum-ionising particle.

The tau jet must contain one, two or three tracks reconstructed in the central tracking chamber. After having removed the tracks and calorimetric energy depositions associated with the identied tau jet, the remaining tracks and calorimetric clusters are grouped into two hadronic jets using the Durham jet algorithm. The tau jet must be separated by at least 25 degrees from the two hadronic jets. For events with a transverse energy imbalance less than 25 GeV and with

cos

miss

0

55 for the polar angle of the missing momentum vector, the angular opening of the tau jet must be smaller than eight degrees. The invariant mass of the system of the two hadronic jets must be larger than 60 GeV and smaller than 100 GeV. The invariant mass of the system of the tau jet and the missing four-momentum must be larger than 55 GeV.

The distributions of the number of tracks reconstructed in the central tracking chamber and associated with the tau jet and of the invariant mass of the two hadronic jets are shown in Figure 4.

5

The above cuts select three events in the data. The selection eciencies and the background contributions are listed in Table 1. The accepted background cross section is dominated by the process e

e

(

).

The dominant systematic error on the signal cross section arises from the uncertainty in the accepted

(

) background cross section which is dominated by nite Monte Carlo statistics and leads to a relative error of 19% on the signal cross section. Systematic errors in the tau-jet identication are derived from a comparison of data versus Monte Carlo using e

e

(

) events as a control sample. Systematic errors on eciencies and accepted background cross sections are derived by comparing dierent Monte Carlo event generators and Monte Carlo samples simulated with dierent W masses and detector energy scales. A total systematic error of 20% on the measured cross section is assigned.

3.4 e +

e !``()

Event Selection

The event selection for the process e

e

(

) requires two leptons and missing energy due to the neutrinos. Low-multiplicity leptonic nal states are selected by requiring between one and six tracks in the central tracking chamber and less than 15 calorimetric clusters. The visible energy of the event is required to be larger than 2% and smaller than 80% of

.

Charged leptons are identied inside the polar angular range of

cos

0

92. For electrons and muons, the lepton identication as described above is applied. For muons not reconstructed in the muon chambers, their minimum-ionising-particle (MIP) signature in the calorimeters is used for identication. Final states from hadronic tau decays are identied as geometrical jets which are reconstructed based on a clustering inside a cone of 30 degrees half-opening angle. At least one identied electron or muon with an energy between 20 GeV and 70 GeV is required.

The selection criteria depend on whether one or two electrons or muons are identied, referred to in the following as lepton-jet and lepton-lepton class.

In the lepton-lepton class, the energy of the second lepton must be larger than 8 GeV and smaller than 70 GeV. In order to reject

(

) events, the acoplanarity between the two leptons is required to be larger than eight degrees. Exactly two tracks must be reconstructed in the central tracking chamber. The transverse energy imbalance must be at least 8 GeV and larger than 10% of the visible energy. In order to reject radiative

(

) events where the photon escapes along the beam pipe, the polar angle of the missing momentum vector must neither point to the beam axis,

cos

miss

0

96, nor to the gap between the barrel and endcap electromagnetic calorimeter. The calorimetric energy not associated with the leptons is required to be less than 10 GeV, and the sum of the energies of jets with

cos

jet

0

95 must be less than 5 GeV.

In the lepton-jet class, a jet with more than 8 GeV energy is required. In order to reject

` +

`

(

) events the acoplanarity between the lepton and the jet as well as between the lepton and any track in the central tracking chamber must be larger than eight degrees. At least one track is required to have a momentum larger than 2 GeV. The missing transverse energy must exceed 20% of the visible energy. Since for muons identied by their MIP signature the momentum resolution is worse, the missing energy vector is required to point at least 23 degrees in polar angle away from the MIP muon. Events containing photons with an energy of more than 10 GeV are rejected.

6

The distributions of the acoplanarity between the two charged leptons and of the energy of the identied electron or muon with highest energy are shown in Figure 5.

Cross Section

The above selection cuts select two events in the data, one electron-muon event in the lepton- lepton class, and one muon-tau event in the lepton-jet class. The combined selection eciencies and the background contributions are listed in Table 1. The signal eciency is determined within the following cuts:

cos

0

96 for both charged leptons, with energies larger than 15 GeV and 5 GeV. The accepted background cross section is dominated by e

e (

) and

+

(

) events.

The dominant systematic error on the signal cross section arises from the uncertainty of 6 fb in the accepted

(

) background cross section due to nite Monte Carlo statistics.

Systematic errors on the lepton identication are derived from a comparison of data versus Monte Carlo using e

e

(

) events as a control sample. Systematic errors on eciencies and accepted background cross sections are derived by comparing dierent Monte Carlo event generators and Monte Carlo samples simulated with dierent W masses and detector energy scales. A total systematic error of 5% on the measured cross section is assigned.

3.5 e +

e !qqqq()

Event Selection

The event selection for the process e

e

(

) requires a four-jet signature, with kinemat- ics compatible with a WW intermediate state. The main background arises from the process e

e

(

), which can lead to multi-jet nal states through gluon radiation and jet recon- struction and has a total cross section about two orders of magnitude larger than the expected signal.

Events with high multiplicity and no missing energy are selected by requiring at least ve tracks in the central tracking chamber, at least 30 calorimetric clusters and a visible energy larger than 0

65

. Requiring the longitudinal energy imbalance normalised to the visible energy to be smaller than 0.25 and rejecting events which contain an electromagnetic cluster with an energy of more than 30 GeV suppresses

(

) events with hard initial-state radiation.

Selected events are clustered with a variable jet-resolution parameter such that four jets are formed. The jet resolution parameter at which the event changes from a four-jet to a three-jet topology,

, must be larger than 0.0025, selecting events with four well separated jets. The determination of the jet energies and angles is improved by a kinematic t imposing four-momentum conservation. This selection accepts 88.4% of the WW

(

) signal while reducing the dominating

(

) background by a factor of 21. A total of 80 events pass this selection.

Two pairs of jets are formed, corresponding to the two W bosons. The chosen jet-jet pairing maximizes the sum of the two jet-jet invariant masses, which yields the correct assignment for about 80% of the selected WW

(

) events.

Because of the very high

(

) background and the similar topology of four-jet events arising in WW and

production, a neural network is used to improve their separation. A three-layer feed-forward neural network [20] with twelve input nodes, one hidden layer with 15 nodes, and one output node is trained on signal and background Monte Carlo such that the output peaks at 1 for the signal, and at 0 for the background. The twelve input variables consist of event

7

shape variables sensitive to the general four-jet topology (

, sphericity, minimal and maximal jet energy, minimal cluster multiplicity of the four jets), to the signal kinematics (sum and dierence of the two W masses and W velocities, maximal acollinearity between jets belonging to the same W), and to the background topology (minimal angle between jets, minimal mass of jets when the event is reconstructed as a two-jet event).

The distributions of the jet resolution parameter

and of the sum of the two jet-jet masses are shown in Figure 6. The distribution of the neural-network output is shown in Figure 7. All selection cuts have been applied. An alternative analysis not based on a neural network yields compatible but less precise results.

Cross Section

The output of the neural network for data events is tted by a linear combination of neural- network output distributions derived from Monte Carlo simulations for signal and background.

A maximum-likelihood t [21] is used to determine the fraction of

(

) signal events in the total sample of selected events. The cross sections of all background processes other than

q^q

(

), corresponding to 2.3 events of the selected 80 events, are xed to their Standard-Model expectations. This allows a determination of the fraction of

(

) events in the accepted sample.

Taking selection eciencies and luminosity into account, the result of the t corresponds to a signal cross section of 0

98

^: _:

pb and a

(

) cross section of 142

pb where the errors are statistical. The measured

(

) cross section is in good agreement with both our dedicated measurements of fermion-pair cross sections and with the Standard Model value [19].

The largest systematic error arises from dierences between data and Monte Carlo distri- butions for the neural-network input variables. Decreasing these dierences by a reweighting procedure changes the signal cross section by 4%, which is taken as a systematic error. This error is dominated by the eect of reweighting

(

) Monte Carlo events as a function of

derived from a comparison of data versus Monte Carlo at 91 GeV centre-of-mass energy. Sys- tematic errors due to the variation of the W mass used in the Monte Carlo simulations and dierent detector energy scales are estimated to be less than 3%. The eect of an imperfect simulation of cluster multiplicities is estimated to be less than 2%. A total systematic error of 5% on the measured cross section is assigned.

4 Results

Signal Cross Sections

The cross sections,

i , of the signal processes

are determined simultaneously in one maximum- likelihood t. The total likelihood is given by the product of Poissonian probabilities,

(

i

i ), corresponding to the signal processes

having

i selected events (Table 2). The expected number of events for process

,

i , is calculated as:

i =

0

@ 5

X

j

ij

j +

_i

1

A

L

i

(1)

where

ij is the eciency of selection

to accept events from process

,

_i

is the remaining background cross section arising from other processes, and

i is the luminosity used in the analysis of process

. These numbers are listed in Tables 1 and 2. For the e

e

(

) process, the Poissonian probability is replaced by the likelihood as a function of the signal

8

cross section derived from the t described in Section 3.5. Statistical errors corresponding to a 68% condence level interval are determined by a change of 0.5 in the logarithm of the total likelihood. The resulting cross sections and their statistical errors as given by the t are listed in Table 2.

For the

e

(

) and

(

) nal state the measured cross sections contain signicant con- tributions from processes not mediated by resonant W-pair production. In order to determine W-pair cross sections also for these nal states the measured cross sections are scaled by a multiplicative factor,

i . These conversion factors are given by the ratio of the total CC03 cross section and the four-fermion cross section within cuts, and are calculated within the Standard Model using the EXCALIBUR [14] event generator. They are determined to be 1.27 for

e

(

) and 0.92 for

(

), where the dependence of the

i on M

is negligible. These cross sections for the

e

(

) and

(

) nal states are also listed in Table 2.

W-Pair Cross Section and W-Decay Branching Fractions

For the determination of the total CC03 production cross section of W-pairs,

, the ansatz described above is modied. The channel cross sections

i are replaced by the product

i

or

i

i for the

e

(

) and

(

) nal states. The ratio

i is the ratio between the CC03 cross section for process

and

. They are given in terms of the W-decay branching fractions,

(W

) and

(W

), as follows:

qqqq = [

(W

)]

,

qq` = 2

(W

qq

)

(W

), and

`` = [1

(W

)]

, where the sum of the hadronic and the three leptonic branching fractions is constrained to be unity.

The total W-pair cross section and the W-decay branching fractions as determined from ts to the data are listed in Table 3. They are determined both with and without the assumption of charged-current lepton universality in W decays. The W-decay branching fractions obtained for the individual leptons are in agreement with each other. This is the rst direct determination of the branching fraction of the W to hadrons. In order to obtain an improved determination of

, the W-decay branching fractions from the Standard Model are imposed, which are calculated including QCD and mass corrections [3] (Table 3). The result for the total production cross section of W-pairs at

= 161

34

0

06 GeV [22] is:

WW

= 2

89

^: _:

(

)

0

14 (

) pb

(2) where the rst error is statistical and the second systematic. This value for

agrees well with other recent measurements of

at

= 161 GeV [23,24].

W Mass

Within the Standard Model the measured cross sections,

i , depend on

and the mass of the W boson, M

. In order to determine a value for M

, the cross-section t to the data is repeated with the cross sections

i of Equation 1 replaced by the functions

i (

M

), leaving M

as the only free parameter. Using the Standard-Model calculations of

i (

M

) as implemented in GENTLE [25] (CC03) and EXCALIBUR [14] (

e

(

) and

(

) nal states), M

is found to be:

M

= 80

80

^: _:

(

)

0

03 (LEP) GeV

The same result for M

is obtained using only the total W-pair cross section of Equation 2, as shown in Figure 8. The second error on M

arises from the uncertainty in the calibration

9

of the LEP beam energy [22]. The error due to the experimental uncertainty of the total W- boson width [26] is negligible. This value for M

agrees well with our indirect determination of M

from measurements at the Z resonance [18], and with recent measurements of M

at pp colliders [5,27] and at LEP [23,24].

Triple-Vector-Boson Couplings

Alternatively, when the W mass is known, the total cross section can be interpreted in terms of triple-vector-boson couplings [28,29]. In particular, it is interesting to test if the coupling between the Z and a pair of W bosons exists [30]. In general, anomalous contributions to W-pair production are parametrised in terms of seven complex triple-vector-boson couplings, too many to be measured with the limited data collected at threshold. Therefore scenarios are considered where a single parameter describes a possible deviation from the couplings predicted by the Standard Model.

Neglecting the contributions of dimension-six operators, assuming that all electromagnetic properties of the W boson are standard and that a SU(2) symmetry is respected leaves a single parameter,

[30]. This parameter describes the deviation of the ZWW coupling,

, from its Standard Model value of cot

= 1

9, where

is the electroweak mixing angle. Our result is:

Z

=

cot

= 0

1

1

9 (68% CL)

= 0

1

3

2 (95% CL)

as shown in Figure 9. Thus our cross section is in good agreement with the Standard Model predictions for the triple-vector-boson couplings and our data favour the existence of the ZWW vertex at about 68% condence level. In an alternative scenario [29], where more than one anomalous coupling is introduced, but depending on a single parameter,

, one nds:

W

= 0

0

0

8 (68% CL)

= 0

0

1

4 (95% CL)

as also shown in Figure 9. This value for

agrees well with other recent measurements of

W

at LEP [24,31].

In order to derive these results with a maximum likelihood t, the GENTLE [25] program is used to calculate the prediction for the total cross sections as a function of the anomalous coupling, using Standard Model W-decay branching fractions and the current world-average value and error for the W mass, 80

33

0

15 GeV [5]. The errors quoted above include the contributions due to the error on the world-average value of the W mass and the systematic errors of the cross section measurements. In both scenarios good agreement with the Standard Model expectation of

=

= 0 is observed. Limits on other anomalous triple-vector-boson couplings are also obtained from measurements at pp colliders [32].

5 Summary and Conclusion

In a data sample corresponding to an integrated luminosity of 11 pb

collected at a centre- of-mass energy of 161.34 GeV, we have measured W-pair production by selecting four-fermion events with high invariant masses. All nal states mediated by W-pair production are analysed.

10

The total W-pair cross section is found to be 2

89

^: :

(

)

0

14 (

) pb. Within the Standard Model, this cross section corresponds to a W-boson mass of 80

80

^: :

(

)

0

03 (LEP) GeV. Alternatively, using an independent determination of the W-mass [5], the cross-section measurement favours the existence of the ZWW vertex; limits on anomalous triple-vector-boson couplings are

3

2 or

1

4 at 95% CL.

6 Acknowledgements

We wish to congratulate the CERN accelerator divisions for the successful upgrade of the LEP machine and to express our gratitude for its good performance. We acknowledge with appreciation the eort of all engineers, technicians and support sta who have participated in the construction and maintenance of this experiment. Those of us who are not from member states thank CERN for its hospitality and help. We also thank D. Schildknecht for useful discussions.

11

References

[1] S. L. Glashow, Nucl. Phys.

(1961) 579;

S. Weinberg, Phys. Rev. Lett.

(1967) 1264;

A. Salam, \Elementary Particle Theory", ed. N. Svartholm, Stockholm, \Almquist and Wiksell" (1968), 367.

[2] D. Bardin et al., Nucl. Phys. (Proc. Suppl.)

(1994) 148;

F.A. Berends et al., Nucl. Phys. (Proc. Suppl.)

(1994) 163.

[3] W. Beenakker et al., in Physics at LEP 2, Report CERN 96-01 (1996), eds G. Altarelli, T. Sjostrand, F. Zwirner, Vol. 1, p. 79.

[4] D. Bardin et al., in Physics at LEP 2, Report CERN 96-01 (1996), eds G. Altarelli, T.

Sjostrand, F. Zwirner, Vol. 2, p. 3.

[5] The UA1 Collaboration, C. Albajar et al., Z. Phys.

(1989) 15;

The UA2 Collaboration, J. Alitti et al., Phys. Lett.

(1990) 150; Phys. Lett.

(1992) 354;

The CDF Collaboration, F. Abe et al., Phys. Rev. Lett.

(1990) 2243; Phys. Rev.

(1991) 2070; Phys. Rev. Lett.

(1995) 11; Phys. Rev.

(1995) 4784;

The D Collaboration, S. Abachi et al., Phys. Rev. Lett.

(1996) 3309.

We use the average value and error for the mass of the W boson as listed in:

R.M. Barnett et al., Review of Particle Properties, Phys. Rev.

(1996) 1.

[6] The L3 Collaboration, B. Adeva et al., Nucl. Instr. and Meth.

(1990) 35.

[7] M. Acciarri et al., Nucl. Instr. and Meth.

(1994) 300.

[8] M. Chemarin et al., Nucl. Instr. and Meth.

(1994) 345.

[9] A. Adam et al., The Forward Muon Detector of L3, CERN-PPE/96-097.

[10] I.C. Brock et al., Nucl. Instr. and Meth.

(1996) 236.

[11] The L3 detector simulation is based on GEANT Version 3.15.

R. Brun et al., GEANT 3, CERN-DD/EE/84-1 (Revised), 1987.

The GHEISHA program (H. Fesefeldt, RWTH Aachen Report PITHA 85/02 (1985)) is used to simulate hadronic interactions.

[12] S. Catani et al., Phys. Lett.

(1991) 432.

[13] The KORALW version 1.21 is used.

M. Skrzypek, S. Jadach, W. Placzek and Z. Was, Comp. Phys. Comm.

(1996) 216;

M. Skrzypek, S. Jadach, M. Martinez, W. Placzek and Z. Was, Phys. Lett.

(1996) 289.

[14] F.A. Berends, R. Kleiss and R. Pittau, Nucl. Phys.

(1994) 308; Nucl. Phys.

(1994) 344; Nucl. Phys. (Proc. Suppl.)

(1994) 163; Phys. Lett.

(1994) 490;

R. Kleiss and R. Pittau, Comp. Phys. Comm.

(1994) 141.

12

[15] T. Sjostrand, PYTHIA 5.7 and JETSET 7.4 Physics and Manual, CERN-TH/7112/93 (1993), revised August 1995;

T. Sjostrand, Comp. Phys. Comm.

(1994) 74.

[16] The KORALZ version 4.01 is used.

S. Jadach, B. F. L. Ward and Z. Was, Comp. Phys. Comm.

(1994) 503.

[17] J. H. Field, Phys. Lett.

(1994) 432;

J. H. Field and T. Riemann, Comp. Phys. Comm.

(1996) 53.

[18] The L3 Collaboration, M. Acciarri et al., Z. Phys.

(1994) 551.

[19] The L3 Collaboration, Measurement of Hadron and Lepton-Pair Production at 161 GeV

<

p

s<

172 GeV at LEP, in preparation.

[20] C. Peterson et al., Comp. Phys. Comm.

(1994) 185.

[21] R. Barlow, J. Comp. Phys.

(1987) 202;

R. Barlow and Ch. Beeston, Comp. Phys. Comm.

(1993) 219.

[22] The working group on LEP energy, CERN-SL note, in preparation.

[23] The OPAL Collaboration, K. Ackersta et al., Phys. Lett.

(1996) 416.

[24] The DELPHI Collaboration, P. Abreu et al., Measurement and interpretation of the W-pair cross-section in

interactions at 161 GeV, CERN-PPE/97-09.

[25] The GENTLE version 2.0 is used.

D. Bardin et al., GENTLE/4fan v. 2.0: A Program for the Semi-Analytic Calculation of Predictions for the Process

4

, DESY 96-233, hep-ph/9612409.

[26] The UA1 Collaboration, C. Albajar et al., Phys. Lett.

(1991) 503;

The UA2 Collaboration, J. Alitti et al., Phys. Lett.

(1992) 365;

The CDF Collaboration, F. Abe et al., Phys. Rev. Lett.

(1995) 341; Phys. Rev.

(1995) 2624;

The D Collaboration, S .Abachi et al., Phys. Rev. Lett.

(1995) 1456.

We use the average value and error for the total width of the W boson as listed in:

R.M. Barnett et al., Review of Particle Properties, Phys. Rev.

(1996) 1.

[27] M. Rijssenbeek, Measurements of the Mass of the W Boson from CDF/D, to appear in the proceedings of the 28th International Conference on High Energy Physics, 25-31 July 1996, Warsaw, Poland.

[28] K. Hagiwara et al., Nucl. Phys.

(1987) 253.

[29] G. Gounaris et al., in Physics at LEP 2, Report CERN 96-01 (1996), eds G. Altarelli, T.

Sjostrand, F. Zwirner, Vol. 1, p. 525.

[30] I. Kuss and D. Schildknecht, Phys. Lett.

(1996) 470.

[31] The OPAL Collaboration, K. Ackersta et al., Measurement of the triple gauge boson cou- pling

from

production in

collisions at

= 161 GeV, CERN-PPE/97- 04.

13

[32] The UA2 Collaboration, J. Alitti et al., Phys. Lett.

(1992) 194;

The CDF Collaboration, F. Abe et al., Phys. Rev. Lett.

(1995) 1017;

The D Collaboration, S .Abachi et al., Phys. Rev. Lett.

(1995) 1023; Phys. Rev. Lett.

75

(1995) 1034; Phys. Rev. Lett.

(1996) 3303.

14